Images of Westpac\u2019s CEO and a high-profile financial journalist were used in financial scams. Meta didn\u2019t respond to one and platforms didn\u2019t resolve the problem for the other. Will the government act?<\/p>\n
It was quite the scene: Westpac chief executive Catherine McGrath was mid-argument with Winston Peters, a TV producer was trying to separate them, and Jack Tame was attempting to calm Peters. No wonder when McGrath\u2019s colleagues saw the AI image circulating on Facebook as the hook for an investment scam, some thought it was a joke. That it was April 1 didn\u2019t help.<\/p>\n
\u201cSome people who knew me well forwarded it saying they thought it was an April Fool\u2019s,\u201d says McGrath. But, of course, it wasn\u2019t a joke \u2013 it was a deepfake designed to funnel New Zealanders toward a fraudulent investment platform.\u00a0<\/p>\n
McGrath did everything she could to get the ad removed, but no matter how many ways she contacted Meta \u2013 the owner of Facebook, Instagram, Threads and WhatsApp \u2013 she never received a response.\u00a0<\/p>\n
If Meta won\u2019t respond to the chief executive of Westpac, who, in New Zealand, will it respond to? Not The Spinoff, apparently. We contacted the company for this story but heard nothing back.<\/p>\n
McGrath says the scam ad eventually disappeared \u2013 possibly after the Financial Markets Authority (who she\u2019d also alerted to the ads) intervened \u2013 but she decided she needed to speak publicly. \u201cThe media rightly put pressure on banks about whether they were doing enough to stop scams,\u201d she says. \u201cI\u2019d love to see the media put the same pressure on social media platforms.\u201d<\/p>\n
\u00a0<\/p>\n
The AI-generated image featuring Westpac\u2019s Catherine McGrath certainly had a lot going on. (Image: Supplied)<\/p>\n
\u00a0<\/p>\n
She\u2019s not alone. The Financial Markets Authority (FMA) has documented deepfake ads featuring prime minister Christopher Luxon \u2013 one Taranaki grandmother lost $224,000 after viewing a fake Luxon video promoting Bitcoin. Politicians, media personalities, financial commentators, and business leaders have all been impersonated via fake pages and accounts on social media using deepfakes to promote fraudulent investment groups on WhatsApp, Signal, or Discord. Kiwibank CEO Steve Jurkovich and deputy prime Minister Winston Peters have appeared in recent scam ads. The FMA identified 110 such ads in a single 24-hour period on Meta platforms.<\/p>\n
Frances Cook, personal finance journalist and host of the Making Cents podcast, has had her image stolen repeatedly. She hasn\u2019t seen deepfakes like McGrath\u2019s but has lost count of fake accounts using her name and likeness.\u00a0<\/p>\n
Before his recent death, well-known psychologist and TV presenter Nigel Latta interviewed people who had been scammed by someone pretending to be her. \u201cIt\u2019s horrible to think people believe I took their money,\u201d she says.\u00a0<\/p>\n
How did platforms respond to Cook\u2019s situation? She says after years of reporting fake accounts with little result, she\u2019s largely given up. \u201cI feel helpless and hopeless. The platforms don\u2019t do anything.\u201d<\/p>\n
The scams are especially effective, Cook says, because some of the impersonations are almost indistinguishable from legitimate work \u2013 a private group offering investment tips feels so on-brand for a financial journalist (although Cook insists that she never offers specific investment advice).\u00a0<\/p>\n
McGrath feels the violation as much as Cook. \u201cIf it were just my image in an AI story, that could be weird and funny. But it was used for an investment scam. You never want your image used to persuade people to part with their hard-earned money.\u201d\u00a0<\/p>\n
She\u2019s acutely aware of her privilege. \u201cI have more than one method to try to get the ad taken down. If I weren\u2019t the CEO of Westpac, I still don\u2019t know if it would have been removed. That\u2019s not good enough.\u201d<\/p>\n
McGrath wants Meta to verify that anyone buying financial-services ads is a legitimate provider. She wants faster action on fraud reports and a closed reporting loop \u2013 in other words acknowledgment that action has been taken. She also wants a direct channel for bank security professionals to escalate validated scam activity. She also thinks Meta should be checking who it is taking advertising money from. \u201cAsking someone to verify that if they\u2019re taking advertising dollars, it\u2019s from a legitimate advertiser is not an onerous step.\u201d<\/p>\n
Basically, at a minimum, she wants New Zealand to have the fraud-prevention tools Meta has already deployed elsewhere.\u00a0<\/p>\n
Since February 2025, Meta has required financial services advertisers targeting Australian users to provide their Australian Financial Services Licence number and run a visible disclaimer on each ad. The tech giant has also partnered with the Australian Financial Crimes Exchange on a Fraud Intelligence Reciprocal Exchange, which has enabled the removal of over 9,000 spam pages and 8,000 AI-generated celebrity impersonation pages in two months.\u00a0<\/p>\n
That direct intelligence-sharing channel is exactly what McGrath is asking for. \u201cWhy should New Zealanders be treated less well than customers in other parts of the world?\u201d<\/p>\n
And yet, those in other parts of the world aren\u2019t happy either. Last week, The Consumer Federation of America took legal action in the US over scam ads, alleging Meta knowingly took steps and adopted policies to make money at the expense of users\u2019 safety. Meta denies the allegations, saying they misrepresent the reality of the company\u2019s work and that it aggressively combats scams on its platforms.<\/p>\n
The class action complaint came after international news agency Reuters published an investigative series<\/a> on Meta, in which it reviewed internal Meta documents that it said showed the company projected it would earn around US$7bn in 2025 from \u201chigher-risk\u201d scam ads. Advertisers below the threshold for deletion but suspected of fraud were charged higher rates rather than removed, Reuters said.<\/p>\n One document set the maximum revenue Meta would forgo to crack down on suspicious advertisers at 0.15% of revenue or US$135m \u2013 against a potentially US$16bn problem. The documents also described a \u201cglobal playbook\u201d in which Meta cleaned up its public ad library \u2013 a tool regulators use to search for fraudulent ads \u2013 by running regulator search terms to remove flagged content, barely denting the overall volume.<\/p>\n Meta disputed Reuters\u2019 reporting, calling the calculations \u201crough and overly inclusive\u201d.\u00a0<\/p>\n According to New Zealand writer and filmmaker Dylan Reeve, the technology to identify scam ads is not the issue. \u201cThey have the tools in-house to catch these. And they weren\u2019t doing it.\u201d\u00a0<\/p>\n